Epidemiology/Health Services Research

B R I E F R E P O R T

Modest Levels of Physical Activity Are

Associated With a Lower Incidence of
Diabetes in a Population With a High
Rate of Obesity
The Strong Heart Family Study
AMANDA M. FRETTS, PHD1 DARREN CALHOUN, PHD2 We excluded SHFS participants who
BARBARA V. HOWARD, PHD2,3 ANDREA M. KRISKA, PHD5 had diabetes at the baseline examination
BARBARA MCKNIGHT, PHD4 KRISTI L. STORTI, PHD5 in 2001–2003 and those without a follow-
GLEN E. DUNCAN, PHD1 DAVID S. SISCOVICK, MD6 up examination in 2007–2009. Addition-
SHIRLEY A.A. BERESFORD, PHD1
ally, we excluded those with a history of
myocardial infarction, stroke, or heart fail-
ure and those who were pregnant, because
OBJECTIVEdTo examine the association of objectively measured participation in low levels
of physical activity with incident type 2 diabetes.
these conditions may influence activity
and diabetes risk. Participants missing
RESEARCH DESIGN AND METHODSdThe study population included participants baseline glucose measures, family infor-
free of diabetes and cardiovascular disease at baseline (n = 1,826) who participated in a follow-up mation, ,18 or $75 years of age, or
examination. Generalized estimating equations were used to examine the association of steps per who had less than 3 days of pedometer
day with incident diabetes. data were also excluded. In total, 1,826
RESULTSdDuring 5 years of follow-up, 243 incident cases of diabetes were identified. When
individuals comprised the study population.
compared with participants in the lowest quartile of steps per day (,3,500 steps), participants in Accusplit AE120 pedometers (Liver-
the upper three quartiles of steps per day had lower odds for diabetes, consistent with a threshold more, CA) were used to measure the
effect. Contrasting the three upper quartiles with the lowest quartile, the odds ratio of diabetes number of steps taken per day (4). These
was 0.71 (95% CI 0.51–0.98). pedometers have known reliability and
validity (5–7). Participants wore the pedom-
CONCLUSIONSdModest levels of physical activity are associated with a lower risk of in- eter on the hip during waking hours for
cident diabetes, compared with lower levels of activity. 7 consecutive days, except while bathing
Diabetes Care 35:1743–1745, 2012
or swimming (8). Incident diabetes was
defined based on 2003 American Diabetes
Association criteria.

G
uidelines recommend accumulating Family Study (SHFS), a population with a The primary measure of physical ac-
10,000 steps per day as part of a high prevalence of obesity, low physical ac- tivity used was the average steps per day
healthy lifestyle (1,2). However, it tivity levels, and a high burden of diabetes. during the time the pedometer was worn.
is not known whether participating in Generalized estimating equations (GEE)
more modest levels of activity is associated RESEARCH DESIGN AND with an independence working correla-
with a lower risk of diabetes in a high-risk, METHODSdThe SHFS is a population- tion structure and robust standard errors
relatively inactive population. The purpose based longitudinal study of cardiovascular were used to examine the association of
of this study was to assess the relationship disease in 13 AI communities in Arizona, pedometer-determined steps per day with
of pedometer-determined steps per day with North Dakota, South Dakota, and Oklahoma. the risk of diabetes. Statistical analyses
incident diabetes among American Indians Details of the study design have been de- were conducted using STATA version 9.0
(AIs) who participated in the Strong Heart scribed previously (3). (Stata, College Station, TX).
We computed the odds ratio and 95%
c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c CI for developing diabetes using GEE.
From the 1Department of Epidemiology, University of Washington, Seattle, Washington; the 2MedStar Health
Odds ratios (95% CI) were calculated
Research Institute, Washington, District of Columbia; the 3Georgetown and Howard Universities Center for using participants in the lowest quartile of
Translational Science, Washington, District of Columbia; the 4Department of Biostatistics, University steps per day as the referent group, after
of Washington, Seattle, Washington; the 5Department of Epidemiology, University of Pittsburgh, Pitts- adjusting for confounding factors.
burgh, Pennsylvania; and the 6Departments of Medicine and Epidemiology, Cardiovascular Health
Research Unit, University of Washington, Seattle, Washington.
Corresponding author: Amanda M. Fretts, amfretts@u.washington.edu. RESULTSdThe study population in-
Received 30 November 2011 and accepted 1 April 2012. cluded 1,149 (62.9%) women, and the
DOI: 10.2337/dc11-2321 median age at baseline examination was
The opinions expressed in this article are those of the authors and do not necessarily reflect the views of the 37.6 years (range 18.7–74.9). The median
Indian Health Service.
© 2012 by the American Diabetes Association. Readers may use this article as long as the work is properly BMI was 32.0 kg/m2 (16.6–68.7). There
cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/ were 178 (9.7%) participants with predi-
licenses/by-nc-nd/3.0/ for details. abetes at baseline.

care.diabetesjournals.org DIABETES CARE, VOLUME 35, AUGUST 2012 1743

Modest activity associated with lower diabetes
Table 1dOdds ratio (95% CI) of type 2 diabetes according to steps per day
Incident diabetes ,3,500
No. of cases 84
Total at risk 479
Age, site, sex adjustment 1.00
Multivariate* 1.00
Additional adjustment for diet↑ 1.00
Additional adjustment for BMI 1.00
*Adjusted for age (years), site, sex, education (years), cigarette smoking (never, ever, current), alcohol use (never, ever, current), and family history of diabetes (%).
↑Adjusting for all variables in multivariate, as well as total calories (kilocalories per day), total fat (% calories), fiber from grains (grams per 1,000 kcal), sweetened
beverage intake (servings per day), fruit/vegetable intake (servings per day). t Because we found no evidence of a dose-response trend, the three upper quartiles were
collapsed to maximize power.
Mean steps per day decreased with
age, and men had higher step counts than
women. The median steps per day for
participants aged ,55 years were 6,696 for
men and 4,770 for women. Among par-
ticipants aged $55 years, median steps
per day were 5,513 for men and 3,452
for women. Moreover, in this population
without major morbidity, physical inac-
tivity was common in all age groups. Ap-
proximately 26% of participants aged
,30 accumulated ,3,500 steps/day,
25% of participants aged 30–49 years ac-
cumulated fewer than 3,500 steps/day,
and 35% of participants aged 50+ accu-
mulated ,3,500 steps/day.
During follow-up, diabetes developed
in 243 participants. Compared with indi-
viduals in the lowest quartile of steps per
day, the odds ratio for diabetes among
those in the steps per day category “3,500–
5,399 steps/day” was 0.76 (95% CI 0.53–
1.11), after adjustment for confounders.
Likewise, the odds ratios comparing
“5,400–7,799 steps/day” and “7,800+
steps/day” with those who accumulated
“,3,500 steps/day” were 0.74 (0.49–
1.13) and 0.77 (0.51–1.16), respectively.
These data are most consistent with a
threshold effect within the first quartile
of steps per day. Contrasting the three up-
per quartiles with the lowest quartile, the
odds ratio for diabetes after adjustment
for confounders was 0.71 (0.51–0.98).
Adjustment for BMI attenuated the odds
ratios (Table 1).
CONCLUSIONSdThe results from
this analysis in an obese and relatively
inactive population indicate that modest
amounts of objectively measured steps per
day are associated with a lower odds of
developing diabetes. The odds ratios for
the association of physical activity with
diabetes were similar among the upper
1744 DIABETES CARE, VOLUME 35, AUGUST 2012
Quartiles (steps per day)
3,500–5,399 5,400–7,799
57 52
465 448
0.74 (0.51–1.08) 0.73 (0.48–1.12)
0.76 (0.53–1.11) 0.74 (0.49–1.13)
0.76 (0.52–1.12) 0.65 (0.43–0.98)
0.95 (0.65–1.39) 0.83 (0.53–1.28)
three steps-per-day quartiles. Participants
who took at least 3,500 steps/day had a
29% lower odds of developing diabetes
compared with more sedentary participants.
We assessed previously the relation-
ship of self-reported participation in
moderate-to-high intensity leisure-time
or occupational activities on the incidence
of diabetes among middle-aged or older
AIs. In the earlier report, when compared
with less active participants, those who
reported any moderate-to-high intensity
leisure-time or occupational activity had a
33% lower odds of incident diabetes (9),
and there was no evidence that more ac-
tivity was better than less activity. Thus,
both studies support a benefit associated
with participation in modest levels of
physical activity on risk of diabetes in an
inactive population.
Several studies suggest that individu-
als who accumulate 10,000 steps per day
have a decreased risk of obesity and better
glucose tolerance compared with individ-
uals who accumulate fewer steps (1,10–13).
However, health benefits do not appear
to be limited to only the most active indi-
viduals; certain health benefits may be
achieved by adding as little as 2,500 steps
per day to baseline activity (1,11,13,14).
Our analyses complement such findings
and indicate that even modest amounts of
activity are associated with lower odds of
developing diabetes.
Observed odds ratios were attenuated
after additional adjustment for baseline
BMI. This attenuation may be because of
confounding by the independent effects of
obesity on physical activity and diabetes
risk or because of the role of obesity as
a mediator. Unfortunately, we cannot differ-
entiate between potential confounding and
mediation by BMI in this analysis. Addi-
tionally, because diabetes in this cohort is
not rare (13% of cohort developed diabetes
7,800+ 3,500+t
50 159
434 1,347
0.74 (0.50–1.11) 0.74 (0.53–1.00)
0.77 (0.51–1.16) 0.76 (0.55–0.99)
0.70 (0.46–1.06) 0.71 (0.51–0.98)
1.04 (0.69–1.59) 0.93 (0.67–1.29)
during follow-up), reported odds ratios
may not accurately estimate hazard ratios.
The results of this study demonstrate
that physical activity of 3,500 or greater
steps per day is associated with a lower
risk of incident diabetes, compared with
lower levels of activity; above 3,500 steps/
day, more was not better than less activity.
This study identifies physical inactivity as
an important factor related to diabetes
and suggests the need for physical activity
education and outreach programs that
target inactive individuals, particularly
AIs who have epidemic rates of obesity,
physical inactivity, and type 2 diabetes.
AcknowledgmentsdThis study was supported
by the I-T32-HL07902 grant and cooperative
agreement grants U01-HL41642, U01-HL41652,
UL01-HL41654, U01-HL65520, and U01-
HL65521 from the National Heart, Lung, and
Blood Institute (Bethesda, MD).
No potential conflicts of interest relevant to
this article were reported.
This work comprises the contribution of
nine authors. A.M.F. was responsible for de-
veloping the research question of interest and
performing the literature review and data anal-
ysis for the project, as well as writing the man-
uscript. B.V.H. and D.S.S. were the senior
investigators on the project. They supervised
all activities and aided in all aspects of the
project, including development of the research
question and writing the manuscript. G.E.D.
and S.A.A.B. participated in all analyses and
writing of the manuscript. B.M. was the bio-
statistician on the project and supervised the
statistical methods of the manuscript as well as
reviewed all drafts of the manuscript. A.M.K.
and K.L.S. were involved in the implemen-
tation of the pedometers used in the study and
reviewed and edited all drafts of the manu-
script. D.C. reviewed and edited all drafts of the
manuscript and helped write the RESULTS and
CONCLUSIONS sections of the manuscript. A.M.F.
is the guarantor of this work and, as such, had
full access to all the data in the study and takes
care.diabetesjournals.org

responsibility for the integrity of the data and
the accuracy of the data analysis.
The authors acknowledge the assistance and
cooperation of the participating tribes, the In-
dian Health Service hospitals and clinics at each
center, and the SHS/SHFS staff.
References
1. Tudor-Locke C, Bassett DR Jr. How many
steps/day are enough? Preliminary pedometer
indices for public health. Sports Med 2004;
34:1–8
2. Tudor-Locke C, et al. How Many Steps/
day are Enough? For Adults. Int J Behav
Nutr Phys Act 2011;8:79
3. North KE, Howard BV, Welty TK, et al.
Genetic and environmental contributions
to cardiovascular disease risk in American
Indians: the strong heart family study. Am
J Epidemiol 2003;157:303–314
4. Marsh AP, Vance RM, Frederick TL,
Hesselmann SA, Rejeski WJ. Objective as-
sessment of activity in older adults at risk
for mobility disability. Med Sci Sports
Exerc 2007;39:1020–1026
care.diabetesjournals.org
5. Schneider PL, Crouter SE, Lukajic O,
Bassett DR Jr. Accuracy and reliability of
10 pedometers for measuring steps over a
400-m walk. Med Sci Sports Exerc 2003;
35:1779–1784
6. Bassett DR Jr, Ainsworth BE, Leggett SR,
et al. Accuracy of five electronic pedom-
eters for measuring distance walked.
Med Sci Sports Exerc 1996;28:1071–
1077
7. Tudor-Locke C, Ainsworth BE, Thompson
RW, Matthews CE. Comparison of pe-
dometer and accelerometer measures of
free-living physical activity. Med Sci Sports
Exerc 2002;34:2045–2051
8. Storti KL, Arena VC, Barmada MM, et al.
Physical activity levels in American-Indian
adults: the Strong Heart Family Study. Am
J Prev Med 2009;37:481–487
9. Fretts AM, Howard BV, Kriska AM, et al.
Physical activity and incident diabetes
in American Indians: the Strong Heart
Study. Am J Epidemiol 2009;170:632–
639
10. Swartz A, Strath S, Bassett D, et al. In-
creasing daily walking improves glucose
DIABETES CARE, VOLUME 35, AUGUST 2012
Fretts and Associates
tolerance in overweight women. Prev Med
2003;37:356–362
11. Sugiura H, Sugiura H, Kajima K, Mirbod
SM, Iwata H, Matsuoka T. Effects of
long-term moderate exercise and in-
crease in number of daily steps on serum
lipids in women: radomized controlled
trial. BMC Womens Health 2002;2:
3–10
12. Tudor-Locke C, Ainsworth BE, Whitt MC,
Thompson RW, Addy CL, Jones DA. The
relationship between pedometer-determined
ambulatory activity and body composition
variables. Int J Obes Relat Metab Disord
2001;25:1571–1578
13. Moreau KL, Degarmo R, Langley J, et al.
Increasing daily walking lowers blood
pressure in postmenopausal women. Med
Sci Sports Exerc 2001;33:1825–1831
14. Tudor-Locke CE, Myers AM, Bell RC,
Harris SB, Wilson Rodger N. Preliminary
outcome evaluation of the First Step
Program: a daily physical activity inter-
vention for individuals with type 2 di-
abetes. Patient Educ Couns 2002;47:
23–28
1745